For some Type-C to DP Adapter, It need reset tcpm state when resume
To avoid some thing wrong with display appear.
Change-Id: Ie23d62ded05cdbc438cd354c5b886ddc45210b42
Signed-off-by: Zhang Yubing <yubing.zhang@rock-chips.com>
move devm_request_irq() to the end of this function to
make sure rcan and napi is initialized.
Change-Id: I160447fe4f45dedffde9c97f1aa1df536baccb77
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
This driver simulate includes not only the display controller but
also the connector and encoder, name it to vkms as how vmwgfx/amd
name this kind of driver.
Change-Id: If56614c41973549e2120912c91569af27820b15d
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
RK3576 supports calculating the exact dclk from the known hclk rate.
Change-Id: I06f3f59118bc6627b9bb92d0e9aedf1848e5c72c
Signed-off-by: Chaoyi Chen <chaoyi.chen@rock-chips.com>
This is extremely useful in diagnosing remote crashes, and is based heavily
on original work by <md@google.com>.
Signed-off-by: San Mehat <san@google.com>
Cc: Michael Davidson <md@google.com>
[ARM] process: Use uber-safe probe_kernel_address() to read mem when dumping.
This prevents the dump from taking pagefaults / external aborts.
Signed-off-by: San Mehat <san@google.com>
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Change-Id: I7586677c4ea0e064b5a38c7972102db3b5771500
The &vop2_data.nr_layers represents the max number of planes that can
be assigned to one VP. Therefore, when checking the max plane number for
VOP2, it may be incorrect to use the &vop2_data.nr_layers.
And the comparison of assigned plane_mask with &vop2_data.plane_mask_base
can already confirm whether every plane has been assigned to one of all
the VPs.
Fixes: 00424a589d ("drm/rockchip: vop2: add support to assign plane mask by default")
Change-Id: I9c2056dea42897150ef155c7e6fd0ad3a43c8132
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
Since &vop2_win.plane_id means the unique plane id based on the
initialization order of planes in vop2_win_init(), it may be better
to rename vop2_plane_id_to_string() to vop2_plane_phys_id_to_string()
based on the definition of &vop2_win.phys_id.
And the movement can help to make more plane related logs more
readable.
Change-Id: If5d362817fc147cbd3a21f67ea44a8c0814899df
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
RK3528/RK3562/RK3576 use the VOP3 architecture, which supports
flexible switching of planes between different VPs. User can switch
planes between different CRTCs based on the &drm_plane.possible_crtcs
in userspace, so the 'PLANE_MASK' property is not required.
Change-Id: Icbd75bc4602fcfd69a10f2be9f5a8375bd7f2abe
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
The RGB interface also supports formats that have multiple cycles per
pixel, such as MEDIA_BUS_FMT_RGB888_3X8, MEDIA_BUS_FMT_RGB565_2X8_LE,
and so on. And the &drm_display_mode.crtc_clock should be multiplied by
the return value of rockchip_drm_get_cycles_per_pixel().
Fixes: 3cb261be89 ("drm/rockchip: drv: fix the dclk calculation of mcu interface")
Change-Id: Idbb6f6ef2ac17dc0dd63d67519733dbd1e29e9dd
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
The RGB interface also supports formats that have multiple cycles per
pixel, such as MEDIA_BUS_FMT_RGB888_3X8, MEDIA_BUS_FMT_RGB565_2X8_LE,
and so on. And the &drm_display_mode.crtc_clock should be multiplied by
the return value of rockchip_drm_get_cycles_per_pixel().
Fixes: 3cb261be89 ("drm/rockchip: drv: fix the dclk calculation of mcu interface")
Change-Id: I85c0168ccfaedbf5566a2c6b8f86e881fbc93ad0
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
VOP sharp/post-scaler/split use the same line buffer.
They must be mutually exclusive, otherwise the picture
will display abnormally.
Change-Id: Ia8e7877826bf2a4484a4060ea4d704a2edc611c3
Signed-off-by: Algea Cao <algea.cao@rock-chips.com>
Issue:
1. On arm32 platform iommu use the mapping->domain, while use the
group->domain on arm64 platform.
2. __iommu_attach_group use group->default to judge whether new device
in kernel 5.10 branch, while use group->domain in kernel 6.1 branch.
For this change, when two devices share the iommu, the first device
attach success, while the second device failed in kernel 6.1 branch.
Method:
1. Set driver_managed_dma=true to ensure that during the IOMMU probe
phase,
even if attach group failed, the mapping will not be released.
2. After the IOMMU probe succeeds, record info->domain=mapping->domain
and
detach it, ensuring that group->domain is set to NULL. This prevents
the
next device which sharing the IOMMU, not return -EBUSY when attach
the
group.
3. Before hardware running, the shared IOMMU should check whether the
device has been switched. If so, it needs to attach the current
device's info->domain.
4. When removing a shared IOMMU, it must also ensure that group->domain
corresponds to the current device.
Change-Id: If0fec01a0bcf9c49850129bfa5ac28484fece9a2
Signed-off-by: Ding Wei <leo.ding@rock-chips.com>
fix the crash caused by interrupt occurring before info->last_chan is ready
Change-Id: I6a57d49d5af6fb35708e009c33422ad936af0d3f
Signed-off-by: Simon Xue <xxm@rock-chips.com>
fix the variable naming conflict with "lock" when CONFIG_ROCKCHIP_SARADC_TEST_CHN enabled
Change-Id: I9f2925fb6e4be642cacf8e5d5c764e33d13b9cc4
Signed-off-by: Simon Xue <xxm@rock-chips.com>
According to the eDP v1.3 chapter 3.6 Table 3-15, Alternative
Scramble Seed Reset(ASSR) is a recommended way for eDP Sink devices
to support Display Authentication and Content Protection as Method
3a, while Method 1 HDCP is normally not expected in an eDP Sink
device.
In addition, the ASSR support capability should be the bit 0 of
DPCD register 0000Dh according to the eDP v1.4 'Revision History'
table 2:
......
Table 3-4: Corrected reference to DPCD Address 0000Dh, bit 0 (was bit 4)
......
Change-Id: Iafad97755fd4c9e688c47164d099b290af905036
Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
